Despite several setbacks and limited clinical success, gene therapeutic treatment of a variety of diseases remains a scientifically sound concept. Recently, AAV-based vectors have emerged as promising candidates for gene therapeutic applications. While AAV vectors have many attractive properties, their ability to transduce a wide variety of cells can be either advantageous or detrimental - depending on the situation. For cell-type specific transduction with AAV, its promiscuity must be eliminated and new targeting information added. Despite significant progress over the last few years, re-targeting of AAV by inserting specific ligands into the capsid remains a difficult task. To overcome this problem we take a 2-prong approach: In 1 approach (Aim 1), we are using linker-insertion mutagenesis to identify the optimal insertion point for specific peptides. Linker-insertion mutagenesis allows the efficient production of libraries that contain viruses composed of capsid proteins with peptide insertions at all possible (and viable) positions. In addition, we will generate AAV-libraries that contain ligands at defined capsid positions but are flanked by a library of linker amino acids. Amplification on target cells and at low multiplicities of infection will allow the identification of the optimal insertion position or linker amino acids. To complement linker-insertion mutagenesis and linker amino acid libraries, we will pursue a second avenue (Aim 2). In this approach, we will produce mosaic viruses - viruses that are composed of both wild type (or receptor binding negative) and mutant capsid proteins. It is expected that this will result in increased infectious titers when compared to viruses that consist of mutant capsids alone. Insertion of peptides into the AAV capsid can result in the destruction of specific capsid functions. These particular functions, however, might not be required to be encoded in all capsid proteins. At the same time, the targeting ligand is most likely not required to be present in all capsid proteins for efficient targeting. As a consequence, mosaic viruses are expected to be more infectious. The goal of Aim 3 is to target long-term repopulating hematopoietic (c-kit positive) stem cells. If successful, this will eventually allow the treatment of a wide variety of hematopoietic disorders. Using a phage display approach, we will identify peptides that bind to c-kit. We will then insert these peptides, as well as kit ligand and a single-chain antibody against c-kit, into the AAV capsid based on the crystal structure and taking advantage of the methods developed in Aim 1 & 2.